Executive Compensation: Research Q&A

Pay for performance is the idea that an executive’s compensation should vary with the performance of the company rather than remain fixed regardless of results. In practice, companies pursue this goal by combining salary with bonuses, stock awards, options, and other incentive compensation tied to financial and nonfinancial measures, including operational and ESG metrics. Most listed firms use some form of incentive based compensation linking executive pay to firm performance in an effort to reduce managerial agency costs. Both sides of the relationship are hard to define. “Pay” may include awards granted in one year, earned over several years, and paid or valued in another. “Performance” may refer to stock return, accounting results, operational targets, or nonfinancial objectives, and a single metric captures only part of the arrangement. A focus on one measure such as total shareholder return disregards the diversity of corporate strategies and the fact that some boards may prioritize the interests of long term shareholders over short term shareholders, or may take into account the interests of nonshareholder constituencies such as employees. The policy question is whether the selected measures, time horizons, and incentives encourage decisions that serve the company’s longer term strategy without creating opportunities for manipulation.

Aligning executive pay with company performance is difficult because neither pay nor performance is a single, stable quantity. Executive compensation can include salary, annual bonuses, multiyear equity awards, pension benefits, severance, sign on awards, and incentives awarded in one period but paid in another. Company performance can be measured through stock return, net income, operational achievements, transactions such as mergers or financings, or nonfinancial objectives. These measures can point in different directions and operate over different time horizons. Total shareholder return is a backwards looking measure and is not useful in assessing how well a company is performing in areas that will determine its long term value and success. Regulatory metrics can also create apparent precision without resolving these problems. The pay versus performance table the SEC adopted in 2022 illustrates the point: it compares “compensation actually paid,” a figure that incorporates changes in the value of equity awards granted in earlier years, with cumulative total shareholder return, so that a single row combines compensation decisions and market movements from different periods. Critics of the 2015 proposal argued that both the “pay” and the “performance” in pay versus performance are too complex as concepts to fit comfortably within a table.

Executive compensation is a corporate governance problem because boards use pay to influence managerial behavior while executives may possess bargaining power in the process that sets their own compensation. Well designed incentives can reduce managerial agency costs by rewarding managers when shareholders benefit. Poorly designed incentives can produce excessive pay, short term decisionmaking, opportunistic changes in compensation or operations, or rewards disconnected from performance. Boards, legal advisers, and executive compensation consultants must work with complex contracts, equity awards, performance measures, tax rules, and disclosure requirements. Shareholders then evaluate the resulting arrangements through lengthy proxy statements and advisory say on pay votes. Federal regulation has responded with increasingly detailed disclosure and governance rules, but the result has been a thicker rulebook, more complex pay structures, and a heavy compliance burden, without much progress toward alignment. By one estimate, the executive compensation section of an S&P 500 proxy statement grew from an average of 5 to 10 pages to 40 to 50 pages between the 1990s and 2015. These rules operate in the complex domains of corporate governance and taxation, where economic actors have strong pecuniary incentives to find ways to undercut or evade them. The persistence of high pay alongside weak performance illustrates why compensation remains both a board responsibility and a recurring target of regulation.

One study examined firms headed by chief executives whose pay ranked in the top 10% in a given year and found that those firms subsequently suffered an 8% drop in stock value in each of the following three years. The authors of that study interpreted the result as consistent with the hypothesis that highly paid CEOs are overconfident and therefore engage in suboptimal behavior and empire building, and that investors overreact to the incentive pay and are subsequently disappointed by the company’s underperformance. The pattern is visible in particular industries as well. In 2014, the median pay of the CEOs of the eleven major media companies was $32.9 million, by far the highest of any industry group in the S&P 500, while the median total return of those companies was only 10.8%, compared with the S&P 500 average of 13.7%. This evidence does not show that incentive compensation is inherently ineffective, but it does show that the slogan “pay for performance” cannot substitute for careful analysis of incentive design. Large equity or bonus opportunities may intensify managerial effort, but they may also magnify overconfidence. The question is how the structure, metrics, and timing of compensation affect behavior and future firm value.

The Dodd-Frank Act introduced a set of executive compensation reforms after the 2008 financial crisis. Four of the most consequential were shareholder say on pay, pay versus performance disclosure, CEO pay ratio disclosure, and mandatory clawback policies. Say on pay requires public companies to hold a shareholder vote on executive compensation and golden parachutes at least once every three years. Pay versus performance disclosure, proposed in 2015 and adopted in August 2022, requires a table showing compensation actually paid to the CEO and other named executive officers alongside total shareholder return, peer group return, net income, and a company selected financial measure. The CEO pay ratio rule, adopted in 2015, requires disclosure of CEO total compensation, median employee total compensation, and the ratio between them. The clawback rule, proposed in 2015 and adopted in October 2022, requires listed companies to maintain policies for recovering erroneously awarded incentive based compensation after an accounting restatement. Dodd-Frank also required disclosure of hedging policies for employees and directors. Together these provisions sought to align executive pay with corporate performance, a redoubling of decades long efforts, and to give shareholders more voice in compensation matters. Assessed while two of the rules were still proposals, the picture was that each reform may appear intuitive in isolation, but together they add complexity and expense and create opportunities for evasion without getting much closer to pay for performance in practice. In May 2026 the SEC proposed to exempt non-accelerated filers from the pay versus performance, pay ratio, and say on pay requirements; that proposal is pending.

Executive compensation rules are vulnerable to evasion because they operate in the complex domains of corporate governance and taxation, where economic actors have strong pecuniary incentives to find ways to undercut or evade them. A rule that burdens one form of pay can induce substitution of another form that falls outside the rule. For example, a clawback covering compensation tied to financial statement measures may lead companies to shift toward fixed salary, to have the board set incentive pay based on a general assessment of performance, or to tie pay to metrics not found in the financial statements, such as store openings, regulatory approvals, or completing a merger, divestiture, restructuring, or financing transaction. If boards want to circumvent the rules, or if executives have enough bargaining power to require clawback free compensation, the rule can be evaded in its entirety by decoupling pay and performance, severing the very link it was meant to strengthen. A pay ratio can be improved without changing CEO pay by outsourcing the work done by the lowest paid workers. A tax rule limiting deductions for salaries above $1 million turned that ceiling into a salary floor while bonus payments took off through complex stock option plans and other devices. Because such redesign preserves economic value while changing the regulatory classification of the payment, policymakers must evaluate likely behavioral responses, not simply the formal requirements of a rule.

Executive compensation reforms can push companies toward conflicting objectives. Companies reduce some aspects of pay in anticipation of the say on pay vote, but according to one study they offset this by increasing other components, with the net effect of increased overall pay. Pay versus performance disclosure highlights total shareholder return, which can lead firms to change their compensation programs to prioritize short term financial performance. Clawback rules make incentive compensation tied to financial reporting measures, including stock price and total shareholder return, more exposed to recovery, encouraging substitution toward salary, discretionary awards, or the strategic and operational targets the rule leaves outside its reach. The pay ratio rule adds pressure based partly on median worker pay; a board controls CEO pay, but median worker pay can be raised more cheaply by redistributing pay toward the median or outsourcing the lowest paid work than by paying workers more. Section 162(m) long denied a deduction for compensation above $1 million unless it was performance based, encouraging performance based pay, while a broad clawback regime makes that same pay less attractive; the 2017 tax act removed the performance pay exception but did not repeal the cap. A board responding to all these rules may create a more complicated package, harder for shareholders to evaluate and easier to manipulate. Regulatory success cannot be assessed rule by rule; what matters is how the combined system changes incentives, compensation design, and corporate behavior.

In Bank and Georgiev’s account, the history of the $1 million deduction limit shows that a simple regulatory ceiling can become a focal point and change the composition of pay without reducing its total amount. In 1993, federal tax rules limited the deductibility of compensation above $1 million paid to the CEO and the next four most highly compensated employees unless the compensation was performance based. The reform was intended to stymie the growth in fixed executive compensation and encourage incentive based compensation. To regulators’ chagrin, the $1 million salary ceiling quickly became the salary floor, while bonus payments to CEOs took off via complex stock option plans and other devices. Instead of providing a fix, the rule exacerbated then existing tensions and precipitated many of the problems that later reforms, including pay versus performance disclosure and clawbacks, have tried to solve. The episode illustrates that economic actors with strong pecuniary incentives will find ways to undercut or evade a rule. When a rule distinguishes sharply between salary and qualifying performance compensation, firms have reason to reclassify or redesign compensation rather than reduce it. Policymakers should examine substitution effects and the interaction with other rules before assuming that a cap or tax penalty will produce its intended result. Rules of this kind can also be difficult to repeal once adopted: the 2017 tax act removed the performance pay exception to Section 162(m) but left the cap in place.

Say on pay is the requirement that public companies hold a shareholder vote on executive compensation and golden parachutes at least once every three years. It is Dodd-Frank’s most visible and innovative pay related provision. The vote is advisory and nonbinding. Boards are not required to abide by or even respond to it, and courts have held that a negative vote does not help overcome the business judgment presumption attached to a board’s decision on executive compensation. The basic design reflects a change of tack: after decades of attempting to align pay with performance chiefly through enhanced disclosure requirements, Congress and the SEC sought help from firms’ shareholders. Shareholders receive extensive compensation information in the proxy statement and can express approval or disapproval through the vote. The concept is intuitive because shareholders stand to gain the most from strong corporate performance and to lose the most from oversized pay or poor performance. In practice, shareholders have been almost guaranteed to support existing pay arrangements: in each of the five proxy seasons between 2011 and 2015, less than 3% of votes at Russell 3000 companies rejected current arrangements, even though the largest proxy adviser, Institutional Shareholder Services, recommended a negative vote for 12 to 14% of companies. Over the same period, only one fifth of investors believed CEO compensation was appropriate in size and structure.

Congress turned to shareholders because decades of attempting to align pay with performance chiefly through enhanced disclosure requirements had not resolved persistent concerns about excessive pay and weak alignment with company performance. Shareholders stand to gain the most from strong corporate performance, and they stand to lose the most from oversized pay or poor performance. In theory, they should have a strong incentive to monitor executive compensation and corporate performance and, if dissatisfied, to vote pay practices down. Say on pay fit a broader effort in Dodd-Frank to give shareholders more voice in matters of executive compensation without transferring the board’s formal authority. The vote is advisory and nonbinding, so boards are not required to abide by or even respond to it. In practice, shareholders have been almost guaranteed to support existing pay arrangements, and some investors make no decision at all, either not voting or relying automatically on proxy adviser recommendations. According to one study, companies reduce some aspects of pay in anticipation of the vote but offset this by increasing other components, with the net effect of increased overall pay. Say on pay may have focused attention on the optics of compensation, but at the cost of increasing the complexity of pay packages and with little evidence that it has aligned pay and performance in meaningful ways.

The first five proxy seasons showed that say on pay votes almost always approved existing compensation arrangements. From 2011 through 2015, fewer than 3% of votes at Russell 3000 companies resulted in rejection in each year. This was a strikingly low failure rate. The result also differed sharply from the recommendations of Institutional Shareholder Services, the largest proxy adviser, which recommended a negative vote at approximately 12 to 14% of the companies holding votes during the same period. The gap suggests that a negative proxy adviser recommendation did not automatically produce a majority vote against compensation. A company could receive substantial opposition and still obtain formal approval. Bank and Georgiev treat these numbers as evidence that shareholders were almost guaranteed to support existing pay arrangements and that say on pay had not lived up to its promise. Even the rare rejection of a pay package could have little or no consequence, because Dodd-Frank made the vote advisory and nonbinding. The data did not support the expectation that shareholders would routinely use say on pay to police the broader disconnect between compensation and performance.

Bank and Georgiev do not explain the gap; it presents it as evidence of how reliably shareholders approve existing pay arrangements. Institutional Shareholder Services, the largest proxy adviser, recommended a negative say on pay vote at roughly 12 to 14% of the companies holding votes in each of the first five proxy seasons, while fewer than 3% of votes at Russell 3000 companies failed in any of those years. The comparison shows that proxy adviser scrutiny alone did not make say on pay a strong enforcement mechanism: most shareholders continued to approve the packages presented to them even where the leading adviser recommended otherwise. The consequences of the rare failure were also limited. Because Dodd-Frank made the vote advisory and nonbinding, boards were not required to abide by or even respond to a negative result, and courts held that a negative vote did not help overcome the business judgment presumption attached to a board’s compensation decision. The evidence suggests that say on pay may have focused attention on the optics of compensation, but at the cost of increasing the complexity of pay packages and with little evidence that it aligned pay and performance in meaningful ways.

A failed say on pay vote does not invalidate the compensation package or legally require the board to revise it. Dodd-Frank made the vote advisory and nonbinding, so boards are not required to abide by or even respond to it, and the board retains its existing authority over executive compensation. A negative vote also has limited significance in shareholder litigation. Courts have held that a negative say on pay vote does not help overcome the business judgment presumption that attaches to a board’s decision on executive compensation. In practice, even the rare rejection of a pay package can have little or no consequence. If anything, the real impact of the say on pay regime has run contrary to the intent of regulators: one study found that companies reduced some aspects of pay in anticipation of the vote but offset those reductions by increasing other components, with the net effect of higher overall pay. Say on pay nonetheless added an element of accountability and investor voice to the federal framework. Congress created a channel for shareholder voice without reallocating decision authority from directors to shareholders, and that same feature made it possible for boards to maintain a package despite formal shareholder disapproval.

Yes. Shareholder scrutiny can alter the components of compensation while leaving the executive’s overall economic package unchanged or even larger. A study found that companies reduced some aspects of pay in anticipation of say on pay votes but offset those reductions by increasing other components of the compensation package, producing a net increase in total pay. Bank and Georgiev do not identify which components were traded off, but the pattern is a form of substitution: a board may remove a feature likely to attract opposition while increasing another, which can improve the package’s voting profile without strengthening the actual relationship between pay and performance. The assessment is that say on pay may have focused attention on the optics of compensation, but at the cost of increasing the complexity of pay packages. The early survey evidence reflected continuing skepticism. Five years after adoption, only about one fifth of investors believed CEO compensation was appropriate in size and structure, and a similarly small share had confidence that pay was clearly linked to performance. Say on pay may affect presentation and design, but there is little evidence that it has done anything to align pay and performance in meaningful ways.

Say on pay shows that creating a formal vote does not ensure frequent opposition or decisive control. Shareholders received extensive compensation disclosure and a recurring opportunity to express a view, yet fewer than 3% of say on pay votes at Russell 3000 companies failed in each of the first five proxy seasons, even though the largest proxy adviser recommended a negative vote at 12 to 14% of companies. The vote’s advisory character preserved board authority and limited its legal effect. A high approval rate can coexist with substantial dissatisfaction. Survey evidence from 2015 indicated that only about one fifth of investors considered CEO compensation appropriate in size and structure, and a similarly small share had confidence that pay was clearly linked to performance. Shareholder voice depends on information. Disclosure informs the votes shareholders are asked to cast, and information intermediaries such as proxy advisory firms analyze and translate that information for investors; the executive compensation disclosure rubric is designed to allow investors to compare CEO pay over time and across the industry. Say on pay supplied a vote but not, on the evidence of its first five years, a meaningful link between pay and performance; the evidence suggests that the main result of Dodd-Frank’s reforms has been a thicker rulebook and more complex pay structures.

Pay versus performance disclosure is designed to show the relationship between compensation actually paid to senior executives and company performance. Section 953(a) of Dodd-Frank directed the SEC to require disclosure of the relationship between executive compensation and the issuer’s financial performance, taking into account any change in the value of its shares, dividends, and distributions. The rule the SEC adopted in August 2022, Item 402(v) of Regulation S-K, requires a table covering five fiscal years that shows, for the CEO and on average for the other named executive officers, total compensation as reported in the Summary Compensation Table and compensation actually paid, alongside the company’s cumulative total shareholder return, the cumulative return of a peer group, net income, and a company selected financial measure. Companies must also list the three to seven financial performance measures they consider most important, and must describe, in narrative or graphic form, the relationship between compensation actually paid and each measure in the table. The intended benefit is transparency: investors should be able to see whether large pay outcomes correspond with results. The difficulty is that a standardized relationship may be less coherent than it appears. Compensation reflects awards made and valued in different years, and a table can omit the operational and nonfinancial measures the board actually used. The comment letter argued that the rule should serve a purely informational function, identify all measures that had a significant role in determining pay, and discourage rather than encourage firm to firm comparisons.

Compensation actually paid, often abbreviated CAP, is the measure of executive pay used in the pay versus performance table. The SEC’s 2015 proposal defined it as total compensation as reported in the Summary Compensation Table, modified to adjust the amounts included for pension benefits and equity awards. The rule adopted in 2022 follows that structure but specifies the adjustments. A company starts with the Summary Compensation Table total, removes the grant date value of equity awards and the change in pension value, and adds back the year end fair value of awards granted during the year, the change in fair value of unvested awards granted in prior years, the fair value at vesting of awards that vested during the year, and the pension service cost for the year. CAP therefore rises and falls with the market value of outstanding awards, and a given year’s figure incorporates compensation decisions made in earlier years. A criticism of the 2015 proposal was that the uniform presentation of such a figure masks the fact that it can be distorted by one off events and incorporates decisions from different time periods, and that capturing the complexity of modern pay practices in a single figure is impossible. One distortion they identified, the doubling up of CAP when a company changes CEOs midyear, was addressed in the final rule, which requires a separate column for each person who served as principal executive officer during the year.

Total shareholder return, or TSR, measures the return to a shareholder from changes in share price plus dividends. Under Item 201(e) of Regulation S-K, it is calculated by dividing the sum of cumulative dividends for the measurement period, assuming dividend reinvestment, and the difference between the share price at the end and the beginning of the period, by the share price at the beginning of the period. The pay versus performance rule uses cumulative TSR, expressed as the value of a fixed $100 investment measured from the beginning of the earliest year in the table, for both the company and its peer group. TSR was a defensible choice: Section 953(a) refers to financial performance taking into account changes in share value and dividends, and registrants already calculate and report TSR under Item 201. Yet TSR is an imperfect measure of financial performance in the pay for performance context. It is backward looking and not useful in assessing how well a company is performing in areas that will determine its long term value and success. Measuring performance through TSR disregards the diversity of corporate strategies and the fact that some boards may prioritize the interests of long term shareholders over short term shareholders, or may take into account the interests of nonshareholder constituencies such as employees. An exclusive focus on the link between pay and TSR may also lead firms to change their compensation programs to prioritize that link, irrespective of the board’s considered judgment about the optimal arrangements for the company.

A midyear CEO change can cause a single annual pay figure to combine exceptional payments to two executives. The departing CEO may receive severance or other one off contractual payments, while the incoming CEO may receive a sign on award. Under the SEC’s 2015 proposal, both amounts would have entered the company’s annual compensation actually paid figure, which Bank and Georgiev described as “doubled up” in the transition year. The resulting total would have appeared unusually high even though the board changed leadership in an effort to align pay with performance, and the table would have suggested the opposite: that the link between pay, measured by CAP, and performance, measured by TSR, was broken. One concern was that one off events of this kind could render historical and peer group comparisons useless or, even worse, misleading. The final rule adopted in 2022 addresses this particular distortion by requiring a separate column, for Summary Compensation Table total and for compensation actually paid, for each person who served as principal executive officer during a covered year, so that a transition year shows each executive’s compensation separately rather than a combined figure. The broader point survives: companies may add narrative clarifying one off events, and the rule requires a description of the relationship between pay and performance, but such explanation does little to promote comparability across companies, and a reader still needs to know which executives and payments produced a number before drawing conclusions about whether the board rewarded or corrected poor performance.

Total shareholder return captures the market return over a selected period, but company performance can extend beyond that window and beyond the stock price. Boards may tie pay to ESG, human capital, or operational measures, or to metrics not found in the financial statements, such as obtaining regulatory approvals, opening a specified number of stores, or completing a merger or restructuring. TSR is also a backwards looking measure and may not capture how well a company is performing in areas that will determine its long term value. An exclusive TSR focus disregards corporate strategy diversity and the fact that some boards may prioritize the interests of long term shareholders over short term shareholders, or take into account the interests of nonshareholder constituencies such as employees. TSR was nonetheless a defensible choice given the prescriptive language of Section 953(a), which refers to changes in the value of the shares of stock and dividends and distributions, and the fact that companies already report TSR under Item 201 of Regulation S-K. The final rule keeps TSR as the anchor but places other financial measures beside it: net income, a company selected measure representing the most important financial measure the company used to link pay to performance, and a tabular list of the three to seven most important financial measures, in which nonfinancial measures may also be included. Georgiev’s position in the rulemaking was that TSR should be one financial measure among several, with disclosure of every measure the board actually relied on.

Depending on their framing, disclosure rules can have behavioral effects: in addition to providing information, they may influence substantive corporate behavior. A rule that prominently compares compensation with total shareholder return may imply that pay for financial performance measured through TSR is the best way to structure an executive compensation program, even though the statute does not require companies to structure pay around that metric. Firms may respond by changing their compensation programs to prioritize that link, irrespective of the board’s considered judgment about the optimal arrangements for the company, and at the expense of linking pay to nonfinancial measures or to financial measures that do not track TSR. The concern is strongest when the table is presented as a rigid, uniform template, because a uniform presentation can legitimate one set of compensation strategies over another across firms with different strategies. There is no evidence that in adopting Section 953(a) Congress sought to encourage companies to link pay to financial performance or to prioritize financial over nonfinancial performance; nothing suggests the rule was meant to fulfill any function other than a purely informational one. The final rule prescribes a uniform table but requires each company to describe the relationship between compensation actually paid and each performance measure in narrative or graphic form, and permits supplemental measures and disclosure so long as they are not presented with greater prominence than the required information. The comment letter urged that companies be encouraged to vary or supplement the table and that disclosure identify the measures that actually influenced pay.

A uniform template has advantages when the underlying information is directly comparable across firms and over time. The pay versus performance rule requires the information in a prescribed tabular format intended to increase transparency and comparability, but the uniform presentation masks the fact that the metrics can be distorted by one off events and incorporate decisions from different time periods. Compensation actually paid adjusts reported compensation for pension benefits and equity awards and, through changes in the fair value of unvested awards, incorporates compensation granted in prior years, while cumulative TSR reflects market movements over the measurement period. TSR can also reflect differences in corporate strategy and in the time horizons boards prioritize. The final rule addressed one distortion identified in the 2015 proposal, the doubling up of a transition year’s figure when a CEO is replaced, by requiring a separate column for each principal executive officer. Companies may add narrative to prevent the table from misleading, and the rule requires a description of the relationship between pay and each measure, but such narrative does little to promote comparability across companies. The comment letter argued that because the information is too noisy, firm to firm comparisons should be discouraged rather than encouraged, and that the tabular format should not become a rigid template. Readers must still ask what each number measures and which decisions and periods it reflects.

Narrative explanation connects the regulatory figures to the company’s actual compensation decisions. A table can show compensation actually paid, total shareholder return, and other measures, but it cannot by itself explain a CEO transition, a sign on award to a new CEO, or the board’s reasons for selecting particular metrics. Narrative can identify which awards came from prior periods, why a company selected a nonfinancial measure, whether the link between a measure and pay was discretionary or automatic, and whether one off events distorted a year. The SEC’s executive compensation rules already recognize this relationship: recognizing that numerical information alone might paint an incomplete picture, the SEC requires firms to present a narrative description of any material factors necessary to an understanding of the information disclosed in the compensation tables. The final pay versus performance rule applies the same logic. It requires a clear description, in narrative or graphic form or both, of the relationship between compensation actually paid and each financial performance measure in the table, and between the company’s TSR and its peer group’s TSR. The existing CD&A discussion, which is prospective and conceptual, does not substitute for an explanation of the link between actual pay and actual performance over the covered period. With such explanation, investors can distinguish temporary anomalies from genuine misalignment and assess the board’s compensation philosophy rather than merely compare isolated data points.

Nonfinancial measures may play a significant role in the compensation program and therefore in explaining why an executive received a particular payout. The statutory mandate adopted in 2010 refers only to financial performance, but since then firms have increasingly considered both financial and nonfinancial performance in setting executive pay, including climate related, human capital, and other ESG metrics as well as operational metrics. A table limited to TSR, net income, and similar accounting measures can present a skewed or incomplete picture when the board deliberately tied compensation to objectives outside those columns. The omission also creates a framing effect by suggesting that financial measures are the legitimate form of performance. Georgiev’s comment on the proposal argued that the disclosure should identify all measures, financial and nonfinancial, that had a significant role in determining pay; that the term “company performance measures” should be defined to cover nonfinancial measures, both hard, numerical ones and softer, qualitative ones; that the specificity of the disclosure should match the specificity with which the measure is incorporated in the plan, with a carve out for commercially sensitive information; and that the disclosure should state whether the link between a measure and pay was discretionary or automatic. The final rule went part of the way. Its table and its required list of three to seven most important measures are confined to financial performance measures, but a company may include nonfinancial measures in the list if they are among its most important, and the relationship description may address them.

One 2021 study reported that more than half of S&P 500 companies had incorporated ESG metrics into incentive-based compensation plans, including both climate-related metrics and metrics related to human capital management. A separate 2021 study of the 100 largest public companies found that 41 incorporated ESG metrics into their executive compensation programs. Among those 41 companies, 33 incorporated ESG metrics into a holistic qualitative review of individual performance, using different weightings and approaches, while 8 included ESG as an individual metric. These figures illustrate both prevalence and heterogeneity: some companies treated ESG as an individual metric, while others placed it within a broader qualitative review, and practices were evolving fast. In addition to ESG metrics, incentive plans also incorporated other nonfinancial metrics, such as operational metrics. The evidence also supports a disclosure design point. A financial table cannot explain a payout if a material part of the decision depended on climate, human capital, operational, or another nonfinancial objective. Nor can a fixed list of financial measures capture rapidly evolving practices. Investors need to know which measures significantly affected compensation and whether the link between a particular measure and pay was discretionary or automatic.

A company performance measure should include any financial or nonfinancial criterion that played a significant role in determining named executive officer pay. The category should not be confined to numbers appearing in the financial statements. Illustrations include total shareholder return, net income, ESG and human capital measures, operational measures, and metrics not found in the financial statements, such as opening a specified number of stores, obtaining regulatory approvals, or completing a merger, divestiture, restructuring, or financing transaction. A measure can be hard and numerical or soft and qualitative. What matters is whether the board used it to assess company performance for compensation purposes. The comment letter argued that the specificity with which a company discloses a measure should match the specificity with which it is incorporated in the incentive plan, with a carve out for commercially sensitive information, and that the disclosure should state whether the link between the measure and pay was discretionary, meaning a non binding factor in the board’s decision, or automatic, meaning a predetermined benchmark in the compensation formula. Defining the category broadly provides a clear historical view of how the board conceptualized and measured company performance and how it linked pay to it. The final rule defines “financial performance measures” to include accounting measures, non-GAAP measures, stock price, and TSR, requires the tabular list to consist of such measures, and permits nonfinancial measures in the list only alongside them; it does not require the discretionary or automatic label.

A fixed number of measures is an arbitrary cut off. The SEC’s 2022 reproposal would have required the five most important performance measures. A company may use fewer than five significant measures, in which case a requirement to identify five would include criteria that did not play a significant role; another may use more than five, in which case the cut off omits part of the compensation logic. The comment letter argued that requiring all measures that had a significant role in determining named executive officer pay would provide a clear historical view of how the board conceptualized and measured company performance for compensation purposes and how it linked executive pay to that performance, and would solve the problem of how to define “most important.” The number of disclosed measures would then follow the actual plan rather than a regulatory template. This information is material to investor voting decisions on director nominations, say on pay, and compensation related shareholder proposals because it allows investors to assess whether the measures fit the company’s strategy. The approach retains a significance threshold, preventing the list from expanding to every minor consideration, and replaces an arbitrary numerical ceiling with a standard tied to the role the measure actually played. The final rule adopted a range rather than a fixed number: the tabular list must contain at least three and no more than seven of the company’s most important financial performance measures, or fewer than three if the company uses fewer.

The same performance measure can operate very differently depending on whether it triggers a formula or informs board judgment. An automatic, non discretionary measure is a predetermined benchmark in the compensation formula. A discretionary measure is a non binding factor in the board’s decision regarding pay, allowing the compensation committee to decide how much weight to give the result and how it interacts with other considerations. Practice among large companies illustrates the range: of the 41 companies among the largest 100 that used ESG metrics in executive compensation in 2021, 33 incorporated them into a holistic qualitative review of individual performance, using different weightings and approaches, while 8 included ESG as an individual metric. A company may state that ESG or human capital performance affects compensation, but the practical significance differs greatly depending on whether the measure is a predetermined benchmark in the formula or a non binding factor in a holistic review. The comment letter therefore urged that the disclosure specify whether the link between a particular measure and pay was discretionary or automatic, and that the specificity of the disclosure match the specificity with which the measure is incorporated in the plan. The final rule did not adopt that requirement, although the required description of the relationship between pay and each measure gives companies room to explain it. The distinction helps investors evaluate how the board exercised discretion and how it linked pay to performance as it understands it, without assuming that every company should use the same type of compensation design.

The SEC’s 2022 reproposal asked whether pretax net income and net income should be added to the table as mandatory measures. Georgiev’s comment argued against it. Mandatory accounting columns would not solve most of the problems associated with total shareholder return, and they would create an additional problem by overemphasizing financial performance. Both measures are already available in the company’s financial statements, so investors who find them relevant to assessing pay for performance can easily obtain them. The point concerns the framing of information and the legitimation of one set of compensation strategies over another, not the availability of information. By listing four financial measures, the table would train investors’ attention firmly on financial performance and privilege management strategies that align pay with short term financial performance over strategies that align pay with a more capacious understanding of performance, including longer term financial performance and nonfinancial or ESG performance. A company selected financial measure would give a company that relies on net income the opportunity to include it, while a company that does not use net income should not be compelled to elevate it. The final rule took a different course: it requires net income as a mandatory column for all covered companies, omits pretax net income, and adds the company selected measure as a further column rather than as an alternative to TSR. The framing concern therefore remains, and the required relationship description is the place where a company can explain that net income played no role in its pay decisions.

A company selected financial measure can give investors a view of performance that better reflects the company’s actual compensation design than a mandatory TSR measure. The comment letter argued that the measure should be presented as an alternative to TSR rather than as an additional mandatory column alongside TSR and net income, since listing four financial measures would train investors’ attention on financial performance and privilege compensation strategies tied to short term financial results. A company might use the field to present pretax net income or net income, if relevant, or TSR calculated over a different time horizon. Some companies do not rely on a single financial measure, while others use several, so a company should be allowed to omit the field when no single measure is used to assess financial performance for compensation purposes, or to identify more than one measure when that is relevant. The measure should be a financial measure, and its selection should not displace disclosure of the nonfinancial measures that also had a significant role in determining executive pay. The final rule requires the company selected measure as an additional column: the financial performance measure not otherwise in the table that, in the company’s assessment, represents the most important measure it used to link compensation actually paid to performance for the most recent year. Smaller reporting companies are exempt from the column, as they are from peer group TSR and the tabular list.

Peer group TSR can be redundant because it is already disclosed elsewhere in the proxy statement under Item 201 of Regulation S-K, and investors who find the information relevant can obtain it easily. Section 953(a) does not require peer group performance information or any other comparative information. Repeating the figure in the pay versus performance table adds prominence rather than new information, and the 2015 proposing release did not state clearly why the column was included. It can also mislead. The column may be taken to imply that incentive based compensation should be a function of the company’s TSR relative to the TSR of its peer group, that is, of its short term financial performance relative to the short term financial performance of its peers. This may make it more difficult to put in place compensation arrangements that link pay to measures of performance other than short term financial performance. Including peer group TSR in the table encourages comparisons based on very noisy information, in an area where such comparisons would not be warranted even if the information were more meaningful. Georgiev’s comment urged that the presentation instead list all performance measures, financial and nonfinancial, that had a significant role in determining executive pay. The final rule retained peer group TSR as a required column for all companies other than smaller reporting companies, using either the Item 201 peer group or the peer group used in the CD&A, and requires a description of the relationship between the company’s TSR and its peer group’s TSR.

The CEO pay ratio rule requires covered public companies to disclose three figures each year: the annual total compensation of the chief executive officer, the annual total compensation of the median employee, and the ratio between the two. Congress enacted the mandate in Section 953(b) of Dodd-Frank in 2010, and the SEC adopted the implementing rule in August 2015. The first company disclosures, from over 3,000 companies, appeared during the 2018 proxy season, and the rule remains in effect. It does not apply to several categories of issuers, including smaller reporting companies, emerging growth companies, and foreign private issuers; a May 2026 SEC proposal would extend the exemption to all non-accelerated filers. Identifying the median employee can be complex, particularly for multinational companies with large, seasonal, part time, or geographically dispersed workforces. The SEC therefore permitted statistical sampling and other reasonable methods, exclusions for employees in jurisdictions with data privacy restrictions and for non-U.S. employees making up less than 5% of the workforce, and use of the same median employee for up to three years absent changes that would significantly affect the ratio. The rule requires only the numbers; it permits but does not require a narrative explanation. These accommodations reduced compliance burdens but weakened reliability and comparability. The ratio appears simple because it condenses two compensation figures into a single number, but its meaning depends heavily on workforce structure, compensation methodology, CEO ownership, and outsourcing.

Compensation ratios have been used in three overlapping ways, each with its own policy dimension, a typology Bank and Georgiev developed. Descriptively, they provide an intuitive way to show how executive and worker pay differ across time and across countries. Estimates of this kind have been used to illustrate the growth of CEO pay relative to that of the average employee, from about 20 times in the 1950s to 120 to 150 times by the early 1990s. Normatively, a ratio can express a judgment about how large the gap should be. Plato argued that the highest-paid should earn no more than five times the income of the lowest-paid; J.P. Morgan suggested that 20:1 would be optimal; Peter Drucker advocated ratios of 25:1 to 30:1 in 1977 and 15:1 to 20:1 in 1984; Ben & Jerry’s and Whole Foods adopted internal caps; and Portland, Israel, and a failed Swiss referendum used specific ratios in legislation. There is, however, no objective benchmark for an optimal pay ratio. Analytically, investors and others have used ratios to analyze individual firms and to compare firms with one another, by analogy to the financial ratios investors routinely use. A firm-level disclosure mandate can be seen as an effort to facilitate this analytical use; the U.S. mandate of 2010 was the first of its kind.

The pay ratio mandate entered Dodd-Frank with little legislative development. Its first appearance was in the Corporate Executive Accountability Act of 2010, introduced by Senator Robert Menendez in February of that year. The provision was later incorporated into Dodd-Frank as Section 953(b). It was not examined by the relevant congressional committees, was not subject to debate, and generated virtually no legislative history. This was notable because Dodd-Frank exceeded 2,300 pages and principally addressed problems related to the 2008 financial crisis, such as too-big-to-fail, excessive risk taking by financial institutions, insufficient coordination among regulators, and inadequate consumer financial protection. The pay ratio appeared in the statute’s executive compensation title alongside say on pay, pay versus performance disclosure, hedging disclosure, and clawbacks. The Senate report spoke only in general terms about giving shareholders a greater voice and more fairness in compensation affairs, and none of Senator Menendez’s statements addressed the specific purpose of the ratio. Postenactment supporters increasingly connected the mandate to income inequality, worker morale, risky behavior, and investor voting. The limited original record left the SEC to identify a securities law rationale during a politically contentious seven year implementation process.

At least three purposes have been attributed to the pay ratio rule. The investor information function treats the ratio as a company specific data point that can help shareholders evaluate executive compensation, cast say on pay votes, or make investment decisions. The behavioral function assumes that public disclosure may pressure boards to reduce CEO pay, increase worker pay, or reconsider internal pay distribution. The public discourse function uses the ratio to focus attention on pay inequity and broader economic inequality. These functions are not mutually exclusive, but they require different kinds of information. Investors need information that is accurate, comprehensible, and complete. A behavioral effect depends on the link between disclosure and the desired change in corporate conduct, which for the pay ratio is attenuated. Public discourse benefits from a simple and salient statistic, but simplicity can conceal methodological weaknesses. The mandate’s sparse legislative history did not clearly choose among these functions. During rulemaking, the SEC emphasized the investor function, stated that the ratio was firm specific and unsuitable for cross company comparison, and disclaimed any intent to “shame” companies. Public and political use after the first reporting season, however, focused heavily on inequality and comparisons. The rule’s effectiveness is best assessed separately against each of the three functions rather than only against the one the SEC invoked.

The rulemaking combined substantial compliance questions with intense political disagreement about executive pay and inequality. Congress specified the basic ratio but left the SEC to determine how multinational companies should identify a median employee, value compensation, treat part time and seasonal workers, handle foreign privacy laws, and permit statistical sampling. One industry estimate placed annual compliance costs at about $710.9 million, while the SEC estimated initial costs as high as $1.1 billion. Public engagement was extraordinary for a technical securities rule. The SEC received more than 287,000 comment letters on the 2013 proposal, most supportive and many generated through form letter campaigns organized by the AFL-CIO and other labor organizations. The provision also faced repeated repeal efforts. The seven year rulemaking process, from enactment in 2010 to the SEC’s final guidance in 2017, reflected this conflict; both the 2013 proposal and the 2015 final rule passed on split three-to-two votes. In 2017, the acting SEC chair reopened discussion by asking companies about unexpected compliance challenges, creating uncertainty even after the final rule. Supporters defended the ratio as useful to shareholders and relevant to inequality, while opponents argued that it would be costly and of little use to investors. The controversy arose partly because the measure was both a securities disclosure rule and a politically resonant symbol of internal pay disparity.

The SEC identified the ratio as a company specific metric that could assist shareholders in evaluating executive compensation practices, and described this as the primary benefit Congress intended. The agency stated that the information might be useful when shareholders cast say on pay votes and could also inform investment decisions. This framing placed the rule within the investor protection tradition of securities disclosure rather than presenting it as a direct response to income inequality. The adopting release did not rely on inequality as the rule’s purpose and emphasized that ratios were not designed for comparisons across companies. That limitation was important because workforce composition, compensation methods, and business models differ widely. The actual reception of the data departed from this regulatory framing. Media organizations, labor groups, think tanks, compensation advisers, and policymakers ranked companies and aggregated ratios, often using the numbers to discuss inequality. Investors and proxy advisers were more cautious: ISS stated that the ratio would not affect its vote recommendations, Glass Lewis said it would not be a determinative factor, and companies that lost say on pay votes in 2018 appear to have done so because of the amount and structure of CEO pay rather than the ratio. Bank and Georgiev conclude that the actual use of the data thus contravened the SEC’s framing of the rule.

Disclosure as soundbite, a term Bank and Georgiev coined, describes a regulatory approach that produces information with high public salience but low informational integrity. The CEO pay ratio is the central illustration. A single number ratio is superficially simple and intuitive, and it can appear to carry far more information than it does. By linking the pay of rank and file workers to that of corporate executives, it takes on a personal dimension that most corporate disclosure lacks. Yet the apparent simplicity conceals differences in workforce composition, outsourcing, geography, part time employment, CEO ownership, partnership distributions, pension valuation, and calculation methodology. The number attracts more attention than its informational content warrants. Conventional disclosure aspires to accuracy, comprehensibility, and completeness. The pay ratio is vulnerable on each dimension: broad inputs and methodological flexibility affect accuracy, the absence of an objective benchmark impairs comprehension, and numbers without sufficient narrative omit necessary context. There is a trade-off between public salience and informational integrity: the features that draw in multiple audiences come at the cost of accuracy, comprehensibility, and completeness. A disclosure rule may achieve unmatched public visibility while proving ineffectual, or even counterproductive, as a means of informing investors or changing corporate behavior. The same pattern appears in other highly prescriptive Dodd-Frank mandates, which generate ambiguous soundbites with low informational value.

The pay ratio condenses a complex compensation system into a familiar comparison: how many times more the CEO earns than the median worker. The result requires little specialized knowledge. It also has a personal dimension because employees and members of the public can compare the median pay figure with their own income. Websites offered interactive tools inviting employees to check how their pay “stacks up with the CEO’s.” The first reporting season shifted the nature of media coverage. Earlier stories generally focused on the sheer size of CEO pay packages, on comparisons among CEOs, and on the disconnect between pay and company performance. Company specific ratios redirected attention toward internal pay disparity and societal inequality. Commentators described the numbers as stark illustrations of inequality, and the New York Times introduced a related “Marx ratio.” Salience can raise public awareness of pay disparities, which the public consistently underestimates, but it also invites overinterpretation. A reported ratio may reflect a seasonal or largely overseas workforce, a one time CEO award, outsourcing, or the compensation measure the company chose; Mattel’s median pay of $6,271, for example, reflected temporary and seasonal employees and a workforce 78% outside the United States. Experimental evidence suggests that laypersons presented with pay ratio data zero in on it and disregard information about firm performance.

Informational integrity, a framework Bank and Georgiev developed, describes the relative extent to which disclosed information exhibits three baseline characteristics of securities disclosure: accuracy, comprehensibility, and completeness. (The Oxford chapter adds relevance as a fourth attribute.) Each is a matter of degree rather than a binary test. Accuracy involves more than the arithmetic: a calculation can follow the rule and still fail to represent the underlying economic reality if its inputs are broad, inconsistent, or selectively chosen. Comprehensibility asks whether users can understand what the number means and whether valid benchmarks exist. Completeness asks whether the disclosure includes the context needed to avoid a misleading impression. The CEO pay ratio illustrates the interaction among these qualities. The reported ratio may be calculated correctly, but different companies can use different methods to identify the median employee and value compensation. Readers may not know whether a high ratio reflects excessive CEO pay, a low wage workforce, global operations, seasonal employment, or an unusual award. The rule’s required numerical disclosure supplies little explanation. Informational integrity is especially important when a metric is publicly salient because audiences may use the number more confidently than its design permits. A sound disclosure regime should evaluate not only whether information is produced, but whether the combination of definitions, methods, presentation, and narrative supports the inferences users are likely to draw.

A ratio can be mathematically correct because the company followed the permitted formula, yet substantively misleading because the inputs do not support the inference a reader draws. Consider a founder CEO who receives a nominal salary while owning stock worth billions of dollars. Alphabet reported a ratio of 0.000005 to one when Larry Page received $1 and the median employee received $197,274, even though his large ownership stake could appreciate substantially. Amazon reported a ratio of 59 to one while Jeff Bezos received reported compensation of about $1.68 million and held approximately 78.9 million shares worth approximately $33 billion in 2017. Partnership distributions can similarly fall outside annual total compensation and produce very low ratios at publicly traded limited partnerships such as Apollo Global Management and the Carlyle Group. At the other extreme, a company with a large part time, seasonal, or low wage overseas workforce may report an enormous ratio. Mattel relied heavily on temporary and seasonal employees, 78% of its workforce was located outside the United States, and its median employee received $6,271; combined with a large new-hire equity award to its CEO, this produced a ratio of 4,987 to one, against 160 to one at Hasbro, which outsources most of its manufacturing. The arithmetic is valid in each case, but the underlying economic realities differ.

The 2018 reporting season revealed enormous variation that reflected company structure as much as compensation policy. For S&P 500 companies, the median reported ratio was about 155 to one. Across the Russell 3000, the median was roughly 70 to one, while one study of more than 2,000 companies reported an average of 144 to one and a median of 69 to one. Industry differences were pronounced. Consumer goods and retail companies, which often employ large numbers of part time and seasonal workers, had a median ratio of 384 to one. Utilities, where employees are more likely to be full time, had a median of 58 to one. Individual outliers were more extreme still: Weight Watchers reported 5,908 to one and Mattel 4,987 to one. Apollo Global Management and Carlyle Group reported one to one, while Twitter, RE/MAX Holdings, and Alphabet reported ratios at or near zero. The SEC itself cautioned that individual ratios are not comparable and should not be aggregated or compared, and the outliers were driven by idiosyncratic features of each company rather than by compensation policy.

Founder CEOs and partnership structures can produce low reported ratios even when the executive’s wealth is enormous. A founder with a large stock position may accept a nominal salary. Alphabet reported a ratio of approximately 0.000005 to one when Larry Page received $1 in compensation and the median employee received $197,274. Amazon reported 59 to one when Jeff Bezos received approximately $1.68 million in reported compensation, although his 78.9 million shares were worth approximately $33 billion in 2017. The ratio captured annual compensation under the SEC rules, not investment returns from dividends and stock appreciation. Berkshire Hathaway reported 1.9 to one on Warren Buffett’s $100,000 salary; counting the $15.1 billion increase in his stock during 2017 would have produced 282,435 to one. A related issue arises at firms organized as publicly traded limited partnerships, such as Apollo Global Management and the Carlyle Group. The CEO is also a partner and receives much of his or her income as distributions from partnership profits, which are not included in the calculation of CEO pay under the pay ratio regulations. Apollo’s CEO received $191.3 million as a partner in the relevant year, none of which entered the ratio. A one to one ratio can thus reflect the legal form of compensation rather than unusually equal pay.

Bank and Georgiev have shown that outsourcing can lower a company’s reported pay ratio without increasing worker compensation or reducing CEO pay. The SEC rule generally counts employees of the reporting company, not workers employed by independent contractors or service providers. Mattel and Hasbro illustrate the point. Mattel makes most of its toys in its own factories, employing 35,820 people in low wage regions of Asia; Hasbro, with 5,400 employees, outsources most production to third party manufacturers, principally in China. The outsourced workers have no effect on Hasbro’s ratio, so Mattel reported 4,987 to one and Hasbro 160 to one, even though the toys are likely made by a similar number of workers earning similar wages in similar locations. Verizon likewise outsources more of its lower paid work than AT&T, which helps explain its higher median worker pay. The rule thus creates an incentive to improve the reported statistic by moving the lowest paid jobs onto a subcontractor’s payroll, even if those workers are paid the same or less. A firm can achieve the same result by redistributing pay from its lowest paid workers to workers at the median without increasing total compensation spending. Narrative disclosure of the structure of the labor force, including reliance on contractors, part time employees, and offshore workers, would make such manipulation harder to conceal.

Workforce composition can dominate the pay ratio. Consumer goods and retail businesses may employ large numbers of part time or seasonal workers whose annual compensation is low because they work fewer hours or only part of the year. The rule permits companies to annualize the pay of permanent employees who worked less than a full year, but not the pay of temporary, seasonal, or part time employees, so a firm with a seasonal workforce can report a much higher ratio than a utility with mainly full time employees. Mattel’s reported ratio reached 4,987 to one and its median employee, a factory worker in Malaysia, earned $6,271; 78% of its workforce was outside the United States, with substantial temporary and seasonal employment. Hasbro, whose employees are generally full time and located in the United States, reported median pay of $74,207 and a ratio of 160 to one. International operations add differences in wage levels and labor markets, and the rule permits, but does not require, cost of living adjustments. The SEC also permits limited exclusions for foreign workers and exemptions where privacy law prevents compliance. A high ratio does not by itself establish excessive CEO pay or poor governance. The SEC has cautioned that ratios are not fit for comparison across firms, and the Mattel and Hasbro figures show why.

Companies used widely different compensation measures to identify the median employee. A 2018 analysis found that 39% used total cash compensation, 28% used base pay and wages, 14% used W-2 wages, and 25% used total direct compensation. The distribution shows substantial methodological variation. AT&T and Verizon illustrate how choices can change the result even among large telecommunications companies: Verizon’s median worker made $127,000 and AT&T’s $78,000, a gap that reflects operational differences in geography, business mix, and outsourcing as well as methodology. AT&T used cash compensation and passed over the first employee identified because that person had an unusual pension benefit. Verizon included company paid health care, estimated company contributions to 401(k) and supplemental retirement plans, and an estimate of the present value of increases in future pension benefits. A separate study found that 16% of companies in its sample included health benefits in total annual compensation. Verizon’s approach therefore captured benefits and required assumptions that AT&T’s approach did not. Each company may have complied with the rule, yet the resulting median pay figures were not calculated on a common basis. This weakens the inference that differences in reported ratios reflect differences in compensation philosophy rather than methodological choice.

Companies made substantial use of the flexibility built into the rule. The de minimis exception permits exclusion of non-U.S. employees when they constitute less than 5% of the workforce. Approximately 24.5% of Russell 3000 companies and 36.7% of all public companies used an exclusion for foreign workers in the first reporting season. The regulations also permit an additional, alternative ratio when it is needed to explain unusual compensation arrangements. One survey found that 13% of companies disclosed an alternative ratio. Common reasons included excluding a one time CEO bonus, pension values, certain equity in a longer term incentive plan, or categories of employees, such as presenting a domestic workforce ratio. Most alternative ratios were lower than the required figure, although several were higher, suggesting that those companies were communicating with their own employees. A separate study of supplemental ratios points in the same direction: only 14% of 1,125 large firms supplied an additional ratio, and 86% of those supplemental ratios were lower than the required ratio. Firms with greater excess CEO compensation and a pay ratio above their industry median were more likely to provide the extra number. Voluntary context can help, but it may also be used strategically, and it appears that the SEC’s reliance on voluntary supplementation was too optimistic.

Aggregate ratios varied because analysts used different samples, dates, definitions, and aggregation methods. During the 2018 reporting season, prominent estimates ranged from 144 to one at the low end to 361 to one at the high end, with several figures in between. An average of company ratios is not the same as a median ratio, and a sample of S&P 500 firms will differ from the Russell 3000 or all reporting companies. Aggregations were also performed at different points in the filing season. Some organizations compared average CEO pay with an estimate of average production and nonsupervisory worker pay rather than aggregating each company’s SEC ratio. The AFL-CIO reported 361 to one, while the Economic Policy Institute continued to use its own methodology and reported 312 to one after criticizing the SEC data. Another study of more than 2,000 companies found an average of 144 to one and a median of 69 to one. Each figure may accurately describe its selected method, but on a superficial level each purported to reflect the economywide ratio. Even where the method of aggregation was flagged, it likely did not register with the public, so the aggregate figures functioned as soundbites available for various rhetorical points.

Before company specific pay ratios became available, annual media coverage typically emphasized the absolute size of CEO compensation, comparisons among highly paid executives, and the disconnect between executive pay and corporate performance. The underlying message was often that CEOs were overpaid or that compensation systems were broken. The 2018 disclosures shifted the focus of coverage. Because each company reported median worker pay and a CEO to worker ratio, coverage focused more directly on the disparity between management and rank and file employees and on the broader problem of income inequality. Many stories drew an express link between high ratios and income inequality, and no disclosure topic other than corporate earnings received comparable national coverage. The Financial Times warned that failure to narrow internal pay gaps could imperil capitalism, and the New York Times introduced a related “Marx ratio” comparing returns to capital and labor. The rule succeeded in shifting public discourse even though the SEC had emphasized investor use and company specific interpretation. This is a central illustration of what Bank and Georgiev call the pay ratio’s public discourse function. Public information can influence debate beyond the audience and purpose identified by the regulator. The shift also magnifies the cost of weak methodology because widely repeated numbers can shape public judgment even when cross company comparisons are unreliable.

Policymakers cited company ratios when proposing taxes, contracting preferences, and other consequences for firms with large internal pay gaps. Portland, Oregon adopted a business tax surcharge for companies whose ratio is 100 to one or higher, and related proposals appeared in California, Illinois, Massachusetts, and at the federal level. Labor unions incorporated company specific rankings into longstanding campaigns about executive pay. The AFL-CIO’s Executive Pay Watch reported an aggregate ratio of 361 to one. The Institute for Policy Studies devoted its 2018 “Executive Excess” report to the new data. The Economic Policy Institute criticized the SEC methodology as diluted by business opposition, retained its own estimate of 312 to one, and supported possible higher tax rates for high ratio firms. Compensation advisers first helped companies prepare for the rule and then produced reports comparing ratios by industry, revenue, workforce size, and other characteristics. At least one performed statistical analysis, and most acknowledged comparability problems but compared the data nonetheless. Their work made the data more visible and easier for others to use. Virtually all of these users compared ratios across firms despite the SEC’s statement that the data was not fit for such comparisons. The variety of responses confirms the ratio’s salience, but also shows how disclosure designed for investors can migrate into advocacy and substantive regulation.

Boards approve CEO pay but do not directly control the compensation of the median employee, so a high ratio can expose the compensation committee to criticism even when its CEO package follows conventional practice. Compensation advisers supplied comparative reports, which suggests some demand from boards, but only five % of companies discussed the ratio in the CD&A section of their 2018 proxy statements, indicating that boards did not formally consider it in setting CEO pay. Proxy advisers took a wait and see approach. Institutional Shareholder Services stated that the ratio would not affect its vote recommendations, Glass Lewis said it would not be a determinative factor, and both reproduced the figures in their reports. The 2018 voting season remained focused mainly on the relationship between executive pay and company performance. Investors did not ignore the data, but its role in investment and voting decisions was uncertain. Employees received a company specific benchmark for the first time and could compare personal pay with the disclosed median, although the median’s meaning depended on methodology and workforce boundaries. Experimental evidence suggests that disclosure of a high ratio reduces consumers’ purchase intentions. Poor comparability matters here because boards and stakeholders may feel pressure from a number that does not isolate the compensation decision they are trying to evaluate.

The United Kingdom adopted a more contextual and standardized approach. Rather than requiring a single comparison with the median employee, the U.K. rule requires listed companies to compare CEO pay with the pay of employees at the 25th, 50th, and 75th percentiles, producing three ratios. The rule also establishes three specified methodologies for identifying those employees and requires the company to state which method it used. The company must also explain whether the CEO to median worker ratio is consistent with its general policies on employee pay, reward, and progression and, if so, why. These features address several weaknesses in the U.S. model. The three percentile points show more of the internal pay distribution. Standard methods improve comparability. The narrative requirement connects the figures to company policy and obliges directors to interpret the result rather than simply publish a number. The United States gives companies greater flexibility and requires less context, which can reduce compliance burden but produces less consistent information. The U.K. design does not eliminate differences in business model or workforce composition, but it is more likely to generate disclosure with stronger accuracy, comprehensibility, and completeness. U.K. companies began reporting under the rule in 2020; Bank and Georgiev’s analysis, written before the first reports, offered the U.K. design as independent confirmation that a numbers only approach is not enough.

The most important improvement would be required narrative disclosure, modeled on the CD&A rules, which follow every executive compensation table with a narrative description of the material factors necessary to understand it. Companies should explain the compensation philosophy applicable to the median worker, the structure of the workforce, including the use of contractors, part time and offshore workers and reliance on automation, and the reasons for significant changes from prior years. Narrative would not make all companies comparable, but it would help readers distinguish executive pay decisions from workforce structure and calculation effects, and it would make it harder for firms to improve their ratio undetected by altering the workforce. Tightening the calculation methodology is not the answer, since its flexibility is what makes the rule workable; mandatory explanation would add little compliance cost because firms already have the information. Voluntary supplements have been selective. Only 14% of a sample of 1,125 large firms supplied supplemental ratios, and 86% of those were lower than the required figure. The SEC had expected companies to add useful explanation, but firms provide context when it improves the presentation and remain silent when it does not. The U.K. rule, with its three percentile ratios, defined methodologies, and narrative requirement, offers independent confirmation that in this context less is not more. A redesigned rule should prioritize informational integrity over the appeal of a simple soundbite.

A compensation clawback requires an executive to return incentive compensation that should not have been received under a specified legal or contractual standard. The Dodd-Frank mandate, implemented by SEC Rule 10D-1 in October 2022 and by exchange listing standards effective in October 2023, focuses on “erroneously awarded” incentive based compensation: compensation received on the basis of a financial reporting measure that an accounting restatement later corrects. Listed companies must maintain recovery policies covering current and former executive officers and compensation received during the three completed fiscal years preceding the date the company was required to prepare the restatement. Recovery is automatic and does not depend on fraud, fault, or the executive’s participation in the error, and the rule reaches both restatements that correct material errors in prior statements and corrections of errors that would be material if left uncorrected in the current period. Sarbanes-Oxley Section 304 was narrower. It applied to the chief executive and chief financial officer, required actual wrongdoing or corporate fraud rather than merely a restatement, and was enforced only by the SEC on a discretionary basis. Sarbanes-Oxley linked recovery to misconduct; Dodd-Frank turned to restitution and strict liability, treating compensation awarded on misstated results as erroneous even if the recipient acted innocently. Critics of the 2015 proposal regarded the rule as overbroad in reaching executives who had no role in the reporting failure, and warned that, like the other Dodd-Frank reforms, it was unlikely to align pay with performance and might undermine the executive compensation regulatory framework the SEC had been building since the 1990s.

Automatic no fault recovery can reach executives whose compensation was affected by a restatement even though they neither caused nor knew about the accounting error. The Dodd-Frank rule covers more than the CEO and CFO. Executive officers include the president, the principal financial officer, the principal accounting officer, any vice president in charge of a principal business unit, division, or function, and any other person who performs a policy making function. The rule reaches current and former executive officers, though only for compensation received after a person began serving as an executive officer, and looks back over the three completed fiscal years preceding the restatement trigger. Once triggered, recovery is required for all covered compensation of all covered executives, irrespective of whether fraud occurred or who was at fault, and companies may not indemnify executives or insure them against the obligation. The only exceptions are narrow: recovery is impracticable when the direct expense paid to a third party to enforce the policy would exceed the amount recovered and the company has made a reasonable attempt at recovery, when recovery would violate home country law in effect before the rule and counsel so opines, or when it would cause a tax qualified retirement plan to fail. This breadth treats the compensation as erroneously awarded rather than as a penalty for misconduct, but it imposes losses without regard to individual responsibility. Critics of the 2015 proposal expected the rule to motivate opportunistic changes in pay arrangements: more fixed salary, incentive pay set by a general board assessment, or metrics outside the financial statements, especially where executives have the bargaining power to demand clawback free compensation. A broad rule can thus recover more money while, in effect severing the link between pay and performance altogether.

Executives and boards can reduce clawback exposure by shifting compensation away from measures the rule reaches. The rule covers incentive based compensation granted, earned, or vested on the attainment of a financial reporting measure, a category that includes accounting measures, stock price, and total shareholder return. It does not reach base salary, awards that vest solely with the passage of time, bonuses awarded at the board’s discretion, or awards based on strategic or operational measures such as completing a transaction or obtaining regulatory approval. The most direct avoidance strategy is therefore more fixed salary and less incentive compensation, although Section 162(m), which denies a deduction for compensation above $1 million paid to covered executives, and a company’s wish to retain some form of incentives may make that unattractive. Incentive awards can also be based on a general assessment of performance by the board rather than a predetermined accounting formula. Another approach is to tie pay to metrics that do not appear in the financial statements. Bank and Georgiev’s illustrations were opening a specified number of stores, obtaining regulatory approvals, or completing a merger, divestiture, restructuring, or financing transaction, and the SEC acknowledged in the proposing release, and confirmed in the final rule, that compensation tied solely to such metrics is not subject to recovery. These achievements may be important, but they can be less clearly connected to the company’s financial performance or stock than the incentive pay they replace. If boards want to circumvent the rules, or if executives have enough bargaining power to require clawback free compensation, the rule can be evaded by decoupling pay and performance. One conclusion was that, instead of strengthening the link between pay and performance, the clawback provision might serve to sever the link altogether.

Clawbacks can change operating decisions as well as compensation contracts. A study examined companies that voluntarily adopted clawback provisions after the Enron and WorldCom scandals in the early 2000s. The study found that the provisions induced companies to make opportunistic operational changes and to focus on the short term at the expense of research and development, substituting real activity for accrual based earnings management. Such actions can improve reported profitability temporarily without improving the underlying business. The study found that any spike in profitability from those measures was reversed within three years, which Bank and Georgiev read as evidence that the changes were a suboptimal shell game designed to evade the clawbacks. If compensation may be recovered after an accounting restatement, executives have stronger reasons to avoid accounting outcomes that could later be challenged, and they may alter production, spending, or investment to meet targets through real decisions rather than through reporting choices. Because there is little evidence that voluntary clawback plans have done much to align pay with performance or to stem the rise in executive compensation, which raises the question whether mandating them is worthwhile. The evidence shows that a clawback rule cannot be judged solely by whether it deters misstatement; it can also change the timing and substance of corporate decisions.

A mandatory clawback regime is likely to escalate the cost of errors in financial statements. When a restatement may trigger recovery from current and former executives, companies have a stronger interest in obtaining additional assurance when faced with difficult accounting judgments. Because of executives’ desire for certainty, audits may become lengthier, which would delay the flow of information to investors and markets. The rule creates corporate law and contract work as well. To ensure that clawbacks are legal and enforceable under the terms of their corporate contracts and foundational documents, companies needed to review and amend bylaw provisions, indemnification provisions, compensation committee charters, and other relevant policies before the December 2023 compliance date, and the final rule prohibits indemnifying executives or paying for insurance against recovery. Foreign law can complicate enforcement: the rule excuses recovery only where it would violate home country law in effect before the rule was adopted, and a foreign company must obtain a legal opinion to that effect. Before treating recovery as impracticable on cost grounds, a company must first make a reasonable attempt at recovery and document it. Companies must also file their policies as exhibits to their annual reports, indicate on the cover page whether the financial statements reflect a correction of an error and whether a recovery analysis was required, and disclose any recovery amounts outstanding. Critics of the proposal regarded these effects as additions to an already heavy compliance burden that offer little prospect of better company management; on that view, the best case is that the requirements lead executives to spend large amounts of money ensuring technical compliance with complex accounting rules.

Executive compensation rules can influence whether a company enters or remains in the U.S. public markets because compliance costs attach to listed status. The clawback rule applies to all companies listed on U.S. exchanges, including foreign private issuers, emerging growth companies, smaller reporting companies, and controlled companies, rather than preserving the exemptions traditionally available to them in executive compensation and corporate governance. The SEC has traditionally allowed foreign private issuers to follow home country rules, but the clawback rule excuses recovery only where it would violate home country law that was in effect before the rule was adopted. Foreign firms may view the additional U.S. requirements as a reason not to list or to delist and rely solely on their home market, particularly as international exchanges become more competitive. Experience after Sarbanes-Oxley showed that many non-U.S. companies withdrew from U.S. markets when faced with complex new governance rules. A U.S. listing confers advantages on non-U.S. companies, but the burden of the clawback rules may outweigh them. Applying the rules to small public U.S. companies raises their compliance costs and could discourage them from going public in the first place. No single mandate need cause a delisting or block an IPO; the concern is that the aggregate volume of obligations added since Sarbanes-Oxley and Dodd-Frank alters the public company bargain, the net benefits of public company status, and strengthens firms’ incentive to avoid the public company ecosystem. A rule designed to improve accountability can thus have a secondary effect on capital formation, one of the core goals of the federal securities laws.